Who needs deepfakes when you can just run a video at reduced speed?
An edited clip of House Speaker Nancy Pelosi, altered to make it look like she’s slurring her words, is making the rounds online following a Wednesday discussion at a Center for American Progress event.
The news comes by way of the Washington Post, which notes that the only digital trickery required to fool millions of Americans into thinking the Democratic congresswoman was drunk in public was to run the video at 75 percent speed and tweak the pitch.
That’s it. No fancy AI or special video editing skills were needed. If you just let out a heavy sigh, you’re not alone.
It’s unclear at present who produced the video, but its murky providence isn’t slowing its success. As of this writing, since being shared Wednesday on the right-wing site PoliticsWatchdog’s Facebook page, the manipulated video has been viewed 2 million times.
It’s also surfaced on various Twitter accounts and other Facebook pages, where many of those sharing the edited film present it as legitimate.
“Nancy is sooo flippin’ drunk… Don’t miss this… Can she be removed for impairment,” read one such tweet.
Google News, Bing News, Yahoo News, 200+ publications
“She obviously has a substance problem,” reads another tweet both linking to the video and tagging Donald Trump’s Twitter account. “Too funny not to watch over and over.”
It’s unclear if those sharing the video realize it’s been manipulated (and are spreading it anyway) or have simply been fooled. While either explanation is possible, the latter is of greater cause for concern. Pelosi has long been a Republican punching bag and this isn’t the first edited video that manipulates her speech — one posted earlier this month to YouTube, which has since been taken down, clocked over 28 million views, the Post reports.
If all it takes to convince people that the Speaker of the House wanders around drunk in public is a slightly edited video, then we’re in serious trouble. If social media denizens can’t spot this low-tech attempt at political manipulation, there’s not much hope they’ll successfully identify more sophisticated efforts in the future.